RocData
Strength analysis of rock and soil masses using the Generalized HoekBrown, Mohr Coulomb, BartonBandis and Power Curve failure criteria
User’s Guide
2004 Rocscience Inc.
Table of Contents
Welcome to RocData! 
3 
Introduction 
4 
Generalized HoekBrown Criterion 
5 
BartonBandis Criterion 
5 
Power Curve Criterion 
6 
MohrCoulomb Criterion 
6 
Summary of Strength Models and Stress Data Types Analyzed 
7 
How Can I Use RocData? 
8 
Determine Strength Parameters 
8 
Generalized HoekBrown Criterion 
8 
BartonBandis Criterion 
8 
Power Curve Criterion 
8 
MohrCoulomb Criterion 
8 
Plot Failure Envelopes 
8 
Estimation of Input Parameters 
8 
Triaxial Lab Test Data 
9 
Triaxial Field Test Data 
9 
Direct Shear Test Data 
9 
Curve Fitting 
9 
Equivalent MohrCoulomb Parameters 
10 
Stress Sampler 
10 
Instantaneous MC Sampler 
11 
Other Generalized HoekBrown Rock Mass Parameters 
11 
Project Settings: Selection of Failure Criterion and Units 
11 
Export Data / Images 
11 
Display Options 
11 
Quick Tour of RocData
12
Entering Input Data 
12 
Interactive Data Input 
12 
Pick Dialogs 
12 
Lab Test Data 
13 
Triaxial Field Test Data 
13 
Viewing of Test Data on Plots 
13 
Estimating Input Parameters for the HoekBrown Criterion 
15 
Estimating Input Parameters for the BartonBandis Criterion 
17 
Estimating Input Parameters for the MohrCoulomb Criterion 
19 
Failure Envelope Plots 
21 
Right Click Menu 
22 
HoekBrown Parameters 
22 
Other Generalized HoekBrown Rock Mass Parameters 
22 
Power Curve Parameters 
23 
MohrCoulomb Parameters 
24 
Equivalent MohrCoulomb Parameters for Curved Failure Envelopes 
24 
Failure Envelope Range 
25 
1
Analysis of Triaxial Test Data 
25 
Analysis of Field (Rock Mass) Triaxial Data 
27 
Analysis of Direct Shear Test Data 
28 
Analysis with Strength Models for fitting both Triaxial and Direct Shear Data 
30 
Fitting of Different Strength Criteria to a Dataset 
30 
Issued Caution when Switching between Data Types 
31 
Selection of Curve Fitting Technique 
31 
Stress Sampler 
33 
Instantaneous MohrCoulomb Sampler 
34 
Exporting Data / Images 
35 
One Click Export to Excel 
35 
Display Options 
37 
Mogi’s Line 
37 
Zoom Extents 
38 
Project Settings 
39 
Documentation 
40 
Tutorials 
41 
Example 1 – HoekBrown Strength Envelope for a Tunnel in Undisturbed
Rock (D = 0) Example 2 – HoekBrown Strength Envelope for a Slope Application in Highly Disturbed Rock (D = 1) Example 3 – HoekBrown, Power Curve and MohrCoulomb Analysis of
41
43
an Intact Rock Triaxial Dataset 
45 
Example 4 – Generalized HoekBrown Analysis of Field Data 
47 
Example 5 – MohrCoulomb and Power Curve Analysis of a Direct Shear
49
Dataset
References
52
2
Welcome to RocData!
RocData is a software program for determining soil and rock mass strength parameters through analysis of laboratory or field triaxial or direct shear data. The program can fit the linear MohrCoulomb strength criterion and three other nonlinear failure criteria – the generalized HoekBrown, BartonBandis and Power Curve strength models – to test data.
RocData includes builtin tables for estimating typical strength parameters for various rock and soil types. This feature, combined with an intuitive interface, rapid response to user inputs, and immediate plotting of strength curves, facilitates parametric studies of strength behaviour.
RocData is a product of Rocscience Inc.
Rocscience produces many other software products for geotechnical, mining and civil engineering applications. Visit www.rocscience.com to find out about the other Rocscience programs.
3
Introduction
Lack of data input on soil and rock mass properties is a major obstacle often encountered in the numerical modeling of geotechnical structures and excavations. The usefulness of elaborate constitutive models, and powerful numerical analysis programs, is greatly limited, if engineers cannot obtain or estimate reliable geotechnical input parameters.
RocData helps to remedy this situation. It provides tools for quickly and easily testing out hypotheses on four of the most widely used and accepted strength models for soils and rock. The program can be used to determine strength models that best describe laboratory or field data.
RocData is designed to aid engineers, especially at the preliminary stages of design. It provides simple and intuitive implementations of the Generalized HoekBrown, BartonBandis, Power Curve and Mohr Coulomb failure criteria. The program enables users to easily visualize the effects of changes in input parameters on rock and soil failure envelopes.
In addition, RocData comes with builtin tables of typical strength parameter values for various rock and soil types, compiled from very credible sources. This allows users to readily obtain reliable estimates of the strength properties of a wide variety of rocks and soils.
The task of determining rock and soil mass properties is usually not an end in itself. It is carried out in order to obtain input material properties for use in limit equilibrium or numerical analysis of geotechnical structures.
The material properties determined from RocData can be used as input for analysis programs such as Phase2 (finite element stress analysis and support design for excavations) and Slide (limit equilibrium slope stability analysis). (Phase2 and Slide are programs also developed by Rocscience.)
There are several different models for describing the strength of materials. In rock and soil mechanics, the most popular models are the Generalized HoekBrown, MohrCoulomb, BartonBandis and Power Curve failure criteria. They are briefly described next.
4
Generalized HoekBrown Criterion This empirical failure criterion establishes the strength of rock in terms of major and minor principal stresses. It predicts strength envelopes that agree well with values determined from laboratory triaxial tests of intact rock, and from observed failures in jointed rock masses.
RocData implements the most recent update (the 2002 edition) of the Generalized HoekBrown criterion. This edition resolves some formerly troublesome issues including:
• The applicability of the criterion to very weak rock masses, and
• The calculation of equivalent MohrCoulomb parameters, from the HoekBrown failure envelope
The Generalized HoekBrown criterion is nonlinear and is expressed as
'
'
σσσ =+ m
1
3
ci
b
σ
'
3
+ s
σ
ci
a
, where
σ
ci
is the uniaxial compressive strength of the intact rock material,
m
b
=
m
s = exp
i
exp
28 14 D
GSI − 100
−
,
GSI − 100 −
9
3
D
, and
a
=
1
1
+
2
6
(
ee
−
−
GSI /15
−
20/ 3
)
.
GSI , known as the Geological Strength Index, relates the failure
criterion to geological observations in the field.
constant for intact rock, while
m i is a material
m
b
and s are rock mass constants.
BartonBandis Criterion The BartonBarton failure criterion is an empirical relationship widely used to model the shear strength of rock discontinuities. It is very useful for fitting a strength model to field or laboratory shear test data of discontinuities. The BartonBarton criterion has the nonlinear form
5
'
τσ
=
'
n
tan
φ
b
+
JRC
log
10
CS
J
σ
'
n
,
where
coefficient and JCS is the joint compressive strength.
φ
b
is the basic friction angle, JRC is the joint roughness
Power Curve Criterion
A substantial amount of experimental evidence suggests that the
failure envelopes of many geotechnical materials, ranging from clays and rockfill to rock discontinuities and rock masses, are not linear,
particularly in the range of small normal stresses. Commonly, the relationship between shear and normal stresses of such curved envelopes can be described with the Power Curve model
'
τ
= a σ + d
n
(
'
)
b
, where
a , b and d are the parameters of the model.
In some literature, the Power Curve is written in the form
'
τ = A P
a
P
'
σ
n
a
+
T
n
,
which involves only nondimensional model parameters A , T and n .
Atmospheric pressure is often used as the normalizing constant
The Power Curve model in RocData can be fit to both triaxial and direct shear data.
P .
a
MohrCoulomb Criterion This is the most common failure criterion encountered in geotechnical engineering. Several geotechnical analysis methods and programs require use of this strength model. The MohrCoulomb criterion describes a linear relationship between normal and shear stresses (or maximum and minimum principal stresses) at failure.
The MohrCoulomb criterion implementation in RocData can be used
to analyze both direct shear and triaxial test data. The MohrCoulomb
criterion for triaxial data is expressed as
σ
'
1
=
''
'
2
c
cos
φ
+
1
+
sin
φ
'
'
1
−
sin
φ
1
−
sin
φ
σ
'
3
, where
c is the cohesive strength, and f is the friction angle.
The direct shear formulation of the criterion is expressed by the equation
τ
'''
'
= c + σφ .
n
tan
6
Summary of Strength Models and Stress Data Types Analyzed
The table below summarizes the strength criteria in RocData and the types of data each can be applied to:
Strength Model 
Type of Stress Data 

Triaxial 
Direct Shear 

Generalized HoekBrown 
X 

BartonBandis 
X 

Power Curve 
X 
X 
MohrCoulomb 
X 
X 
7
How Can I Use RocData?
The following tasks can be accomplished with RocData.
Determine Strength Parameters
Generalized HoekBrown Criterion Determine the Generalized HoekBrown strength parameters of a rock mass (mb, s and a), based on the following input data:
• unconfined compressive strength of intact rock sigci
• the intact rock parameter mi
• the geological strength index GSI
• the disturbance factor D
BartonBandis Criterion Determine the BartonBandis strength parameters JRC and JCS of a discontinuity based on direct shear test results and the following input:
• basic friction angle of the discontinuity fb
Power Curve Criterion Determine the Power Curve strength parameters of a soil or rock (mass) ( a , b and d , or alternatively A, T and n ).
MohrCoulomb Criterion Determine the MohrCoulomb strength parameters, c (cohesive strength), and f (friction angle) of a soil or rock (mass).
Plot Failure Envelopes
Plot the failure envelopes of the four strength models in principal and / or shearnormal stress space.
• Interactively change the parameters of the various strength models to see how they influence the failure envelopes
Estimation of Input Parameters
For the Generalized HoekBrown, BartonBandis and MohrCoulomb failure criteria, conveniently estimate the strength parameters of various rock and soil types using charts and tables built into RocData.
8
Triaxial Lab Test Data
Determine the intact rock parameters (sigci and mi) of the Hoek Brown criterion, the parameters (c and phi) of the MohrCoulomb criterion, or the Power Curve parameters a, b and d (or alternatively A, T and n) from triaxial lab test data.
The triaxial data
• Can be imported from Microsoft Excel through the clipboard, from tabdelimited or comma separated value text files, from RocLab or RocData files, or
• Entered into the program directly using a builtin spreadsheet
Triaxial Field Test Data
For the Generalized HoekBrown criterion, determine the rock mass parameters (mb, s and a) from triaxial field test data on rock mass strength, with the uniaxial compressive strength, sigci, and the disturbance factor, D, as input.
Field data can be imported in the manner described above for triaxial lab data.
Direct Shear Test Data
Determine the parameters (c and phi) of the MohrCoulomb criterion, (JCS and JRC) of the BartonBandis criterion, or (a, b, d or alternatively A, T and n) of the Power Curve model from direct shear test data. For the BartonBandis criterion the basic friction angle, phib, must be supplied as input.
Direct shear data can be imported in the manner described above for triaxial lab data.
Curve Fitting
RocData provides three methods for fitting strength models to test data. The LevenbergMarquardt method is the default technique for fitting all strength criteria to data points. This robust algorithm has become the standard for nonlinear regression. It is very reliable in practice, and has the ability to converge quickly from a wider range of initial guesses than other typical methods.
Users can also fit strength models to data using the Simplex method. The Simplex method is one of the best curve fitting methods, and has a reputation for being very reliable.
9
Linear Regression (linear leastsquares) curve fitting is the third technique provided in RocData. It can be used to only fit the Hoek Brown criterion for intact rock, and the MohrCoulomb strength model to lab data.
The following table summarizes which curve fitting methods are available for the different strength models:
Curve Fitting Method 

Strength Model 
Levenberg 
Simplex 
Linear 
Marquardt 
Regression 

HoekBrown (lab data / intact rock) 
X 
X 
X 
Generalized HoekBrown (rock mass / field data) 
X 
X 

BartonBandis 
X 
X 

Power Curve 
X 
X 

MohrCoulomb 
X 
X 
RocData calculates the sum of the square of the vertical distances of the given data points from a fitted curve, known as “Residuals” in the program. This value is a measure of how well a strength criterion fits a given data set.
Equivalent MohrCoulomb Parameters
For the nonlinear strength criteria (HoekBrown, BartonBandis and Power Curve), calculate equivalent MohrCoulomb strength parameters (cohesion and friction angle).
• The bestfit MohrCoulomb strength envelope is determined over a stress range that you can define based on your application (i.e. tunneling or slope stability).
• Plot the equivalent MohrCoulomb failure envelope in principal and / or shearnormal stress space
Stress Sampler
Graphically sample any failure envelope to determine specific stress (principal, shear or normal stress) values at any point along the envelope.
10
Instantaneous MC Sampler
Graphically determine the instantaneous MohrCoulomb parameters at any point along any of the curved (nonlinear) failure envelopes.
Other Generalized HoekBrown Rock Mass Parameters
RocData calculates rock mass parameters such as tensile strength, uniaxial compressive strength and deformation modulus for the Generalized HoekBrown criterion.
Project Settings: Selection of Failure Criterion and Units
Select a failure criterion or switch between criteria using the Project Settings dialog. In the dialog you can also supply a title for your RocData project.
The Project Settings dialog also allows analyses to be carried out in either Metric or Imperial units. In Metric, the available stress units are Megapascals (MPa), Kilopascals (kPa) and Tonnes per square metre (tsm). In Imperial, the stress units may be either Kilopounds per square foot (ksf), Kilopounds per square inch (ksi), Pounds per square foot (psf), Pounds per square inch (psi) or Tons per square foot (tsf).
Export Data / Images
Export data for further analysis or report writing:
• Copy the data and / or plots to the clipboard for easy import into Microsoft Word or your favourite word processor or imageediting program
• Copy the data and / or plots directly into Microsoft Excel
• Save the plots to a JPEG, BMP, EMF or WMF image file
• Print and Print Preview capabilities.
Display Options
Numerous Display Options to customize the appearance of your plots:
• Change colors, fonts, line thickness.
• Grid overlay.
• Add plot title and show input data directly on plots.
• Zooming.
• Plot Mogi’s line (transition from brittle to ductile failure).
• Grayscale for output to black and white printers
11
Quick Tour of RocData
The following “quick tour” of RocData will familiarize the user with the features of the program.
Entering Input Data
The primary means of user interaction with RocData, is with the sidebar (Docking Form) data input area. The sidebar input area for the Generalized HoekBrown method is shown below. The sidebar is used for data input, and for displaying calculated output parameters.
failure envelopes.
Input data can be entered in various ways:
Interactive Data Input
• You can click on the
arrows with the
mouse to change the input data. All output data is immediately recalculated, and the failure envelope plots are redrawn.
This feature allows you to interactively observe the effect of parameter changes on the shape of failure envelopes, and on computed output values.
Pick Dialogs
• You may also enter input parameters using the “Pick” dialogs. When you select
a Pick button
with a dialog, in the form of a chart or table, which guides you to estimate an appropriate value for the input parameter. Upon selecting OK in the dialog, the selected parameter value will be loaded into the sidebar data input area. RocData proceeds to automatically calculate all output data, and update the
, you will be presented
12
Lab Test Data

• For the Generalized HoekBrown criterion, you may determine values of sigci and mi, from triaxial lab test data for intact rock. This is done with the Use Lab Data option, as described in a later section. When you select Apply or OK in the “Calculate sigci, mi from Lab Data” dialog, the RocData calculation is carried out, and all output values and failure envelopes are updated. The Mohr Coulomb and Power Curve criteria can be also used to fit triaxial lab data. 

• The MohrCoulomb, BartonBandis and Power Curve criteria can as well be used to fit direct shear test data through selection of the Use Lab Data option. 

Triaxial Field Test Data 


• 
Given sigci and D values, the Generalized HoekBrown model can be fit to triaxial field test (rock mass) data. 
Viewing of Test Data on Plots 


• 
When test data is analyzed, the data points can be plotted in RocData’s main view, in addition to the failure envelopes. The data points can be toggled on and off by selecting the Plot Test Data option. For triaxial test data, the lab or field data is plotted as points in principal stress space, and as Mohr circles in direct shear space. For direct shear data, test data is only plotted as points in direct shear space. Examples of such plots of triaxial and direct shear test data are shown below. 
13
Plot of triaxial data on failure envelopes
Plot of direct shear data on failure envelopes
14
Estimating Input Parameters for the HoekBrown Criterion
Each of the parameters used as input for the HoekBrown criterion – sigci, mi, GSI and D – can be estimated using convenient charts and tables built into RocData.
These charts and tables are accessed by selecting the “Pick” button located beside each of the input parameter edit boxes in the sidebar (docking form).
When you select a Pick button, a table or chart will appear, allowing you to determine a suitable value for the desired parameter. For example, the dialogs for estimation of mi and GSI (Rock Type = General) are shown below.
Once you have determined a value, select OK in the Pick dialog. The value will be automatically loaded into the docking form, and the RocData calculation will automatically be carried out (i.e. output parameters and failure envelopes will be recalculated using the selected input data).
15
GSI chart (Rock Type = General)
Select the Pick button for each of sigci, mi, GSI and D, and experiment with the dialogs.
Notice that there are TWO distinct GSI charts:
• One for general rock mass types, and
• A second for weak, heterogeneous rock mass types such as flysch, which extends the useful range of GSI down to values as low as 5.
For further information about the HoekBrown classification parameters sigci, mi, GSI and D, please consult Ref.1.
16
Estimating Input Parameters for the BartonBandis Criterion
Each of the parameters used as input for the BartonBandis criterion – phib, JRC, and JCS – can be estimated using convenient charts and tables built into RocData.
These charts and tables are accessed by selecting the “Pick” button located beside each of the input parameter edit boxes in the sidebar.
When you select a Pick button, a table or chart will appear, allowing you to determine a suitable parameter value. The dialogs for estimating phib and JRC are provided below as examples.
Phib dialog
17
JRC dialog
Once you have determined a value, select OK in the Pick dialog. The value will be automatically loaded into the sidebar input data area, and the RocData calculation automatically carried out.
The material list in the phib dialog can be filtered by moistened state (dry or wet), while the list in the JCS dialog can be filtered into properties for rocks and those for soils.
Select the Pick button for each of phib, JRC, and JCS, and experiment with the dialogs.
18
Estimating Input Parameters for the MohrCoulomb Criterion
MohrCoulomb parameters c and phi can be estimated using RocData’s conveniently built in tables.
These tables are accessed by selecting the “Pick” button
beside each of the input parameter edit boxes in the sidebar.
located
The tables for c and phi are shown below. Using the search option in the dialogs, you can quickly locate the parameters for a material of interest.
Cohesion, c, values for a variety of rock and soil types
19
Friction angle, phi, values for various soil and rock types
Select the Pick button for c and phi to experiment with the dialogs.
20
Failure Envelope Plots
RocData plots failure envelopes in two different spaces:
• Principal stress space (sigma1 vs. sigma3)
• Shear – Normal stress space (sigma normal vs. Tau)
The strength curves plotted correspond to the current data in the docking form. By default, BOTH principal stress AND shearnormal plots are displayed.
Various additional display / analysis options are available, including:
• Display of equivalent MohrCoulomb envelope
• Stress sampler / Instantaneous MohrCoulomb sampler
• Customization of plot appearance with Display Options (eg. grid overlay, line thickness, fonts etc), and zooming
21
Right Click Menu TIP – most of the plot display options, are also conveniently available in the rightclick menu, if you rightclick the mouse anywhere in the plot display area.
HoekBrown Parameters
For a given set of input parameters (sigci, GSI, mi and D), RocData calculates the parameters of the generalized HoekBrown failure criterion (mb, s and a).
These are displayed in the sidebar, as shown below. The Generalized HoekBrown failure envelopes plotted in RocData are generated using the values of mb, s and a shown on the sidebar.
These parameters are calculated using the latest version of the Hoek Brown failure criterion. For a definition of these parameters, and the equations that define them, please see Ref. 1.
Other Generalized HoekBrown Rock Mass Parameters
At the bottom of the sidebar for the Generalized HoekBrown criterion, you will notice that the following rock mass parameters are also calculated:
• Sigt (rock mass tensile strength)
• Sigc (uniaxial rock mass compressive strength)
• Sigcm (global rock mass compressive strength)
• Em (rock mass modulus of deformation)
22
For the definitions of these parameters, and the equations used to calculate them, please see Ref. 1.
In particular, you will notice that there are two calculated values of rock mass compressive strength– sigc and sigcm. For a discussion of how these two parameters are defined, and how they may be used, please see Ref. 1 – section 5: Rock Mass Strength.
Power Curve Parameters
For a given set of input parameters (a, b and d), RocData calculates the parameters of the nondimensional version of the Power Curve criterion (A, T and n), and the tensile and uniaxial compressive strengths (sigt and sigc) of corresponding to the input parameters. These are displayed in the sidebar, as shown below.
23
MohrCoulomb Parameters
For a given set of input parameters (c and phi), RocData calculates the uniaxial compressive strength sigc.
Equivalent MohrCoulomb Parameters for Curved Failure Envelopes
In addition to the parameters of the nonlinear failure criteria (the Generalized HoekBrown, BartonBandis and Power Curve models), RocData always calculates MohrCoulomb parameters (cohesion and friction angle) equivalent to a currently specified strength model.
Since most rock engineering software is still written in terms of the MohrCoulomb failure criterion, the calculation of these equivalent MohrCoulomb parameters is an important feature of RocData.
24
The corresponding MohrCoulomb envelope can be viewed by selecting the MohrCoulomb Envelope option from the toolbar or the Analysis menu.
Failure Envelope Range
It is important to note that the Failure Envelope Range option has
a direct effect on the calculated MohrCoulomb parameters.
• For details about the MohrCoulomb fitting procedure, and the significance of the Failure Envelope Range option, see Ref. 1 (section 4: MohrCoulomb Criterion, and section 6: Determination
of
σ′
3
max
)
• The General option is available only for the Generalized Hoek Brown criterion. When the Failure Envelope Range option is set to this option, notice that sigma3max = sigci / 4. This is based on the empirical observation that the stress range associated with brittle failure, occurs when sigma3 is less than about onequarter of sigci.
• When the Failure Envelope Range option = Custom, you may enter any value of sigma3max or sigNmax.
• The Failure Envelope Range has NO effect on the calculated parameters of a nonlinear failure envelope.
The appropriate stress range (maximum sig3 value) over which equivalent MohrCoulomb parameters must be obtained depends on the application, i.e. the type of geotechnical excavation being designed or analyzed. Based on the principles outlined in Ref. 1, RocData estimates maximum sig3 values for tunneling and slope design applications.
Analysis of Triaxial Test Data
A fundamental feature of RocData, is its ability to input triaxial
stress test data (sigma1 / sigma3 data pairs), in order to determine strength parameters. Three of the failure criteria in the program – the
Generalized HoekBrown, the MohrCoulomb and the Power Curve – allow analysis of triaxial test data. Such analysis is performed as
follows:
Select the Use Lab Data option from the Analysis menu or the toolbar.
1.
2.
3.
The data can be entered in a spreadsheet, or imported from a file, as shown in the dialog below.
A curve fit is then performed on the data, using the Levenberg Marquardt, Simplex or Linear Regression curve fitting techniques,
25
and values of sigci and mi (HoekBrown strength of intact rock data), c and phi (MohrCoulomb strength), or a, b and d (Power Curve strength) are obtained. These calculated strength parameters are known as ‘bestfit’ values.
4. After entering the data, select OK in the dialog, and the ‘bestfit’ values will be used to calculate the failure envelopes and other parameters of the selected strength model.
Example showing fitting of the HoekBrown criterion to intact rock lab data.
Obtaining actual values of strength parameters from triaxial lab test data is always recommended. It should be emphasized that a large number of test results is not always necessary, and good results can be obtained from a relatively small number of data points (eg. 6 or 7 triaxial tests).
If triaxial lab data is not available, parameter values can always be estimated in RocData, using the Pick dialogs described previously. If incorrect data pairs (i.e. data for which sig3 exceeds sig1) are accidentally imported into triaxial data analysis, the curve fitting calculations are not carried out, and an error notice is issued. An example is shown below.
26
Errors issued in dialog when a direct shear data set is incorrectly loaded as triaxial data.
Analysis of Field (Rock Mass) Triaxial Data The Generalized HoekBrown strength model can be fit to triaxial field data. The procedure is the same as above except the Use Rock Mass Data option has to be employed. In addition, the analysis of field data requires entry of intact rock uniaxial strength, sigci, and disturbance factor, D, as input.
27
Fitting of the Generalized HoekBrown criterion to triaxial field test data
Analysis of Direct Shear Test Data
RocData can be used to fit the MohrCoulomb, BartonBandis and Power Curve strength criteria to direct shear (shear–normal) test data. Such analysis is performed as follows:
1. Select the Use Lab Data option from the Analysis menu or the toolbar.
2. The data can be entered in a spreadsheet, or imported from a file, as shown in the dialog below.
3. A curve fit is then performed on the data, using the Levenberg Marquardt, Simplex or Linear Regression curve fitting techniques, and values c and phi (MohrCoulomb strength), JRC and JCS (BartonBandis strength), or a, b and d (Power Curve strength) are obtained. (Fitting of the BartonBandis criterion to direct shear data requires the input of the basic friction angle, phib.)
4. After entering the data, select OK in the dialog, and the computed ‘bestfit’ values will be used to calculate the failure envelopes and other parameters of the selected strength model.
28
Example showing Simplex fitting of the BartonBandis criterion to direct shear data.
29
Analysis with Strength Models for fitting both Triaxial and Direct Shear Data
The MohrCoulomb and Power Curve strength criteria can be fit to both triaxial and direct shear data. At the top of the data analysis dialogs for these criteria, there are two radio buttons for selecting the required data type.
Notice the radio buttons in the upper left corner of the dialog for selecting the type of data to be analyzed for Power Curve strength parameters. In the example shown, the strength model was fit to triaxial data.
Fitting of Different Strength Criteria to a Dataset
RocData allows users to test the fit of different strength models to a given test dataset. Such analysis is performed as follows:
1. Select the Project Settings option from the Analysis menu or the toolbar. Select the desired strength model from the resulting dialog.
2. Next select the Use Lab Data option from the Analysis menu or the toolbar, and enter the test data as previously described.
30
3.
To test the fit of a different strength model to the same data, go to the Project Settings dialog and select the new strength model. Click the OK button in the Project Settings dialog.
4. The data dialog for the selected strength criterion is immediately opened, with the previously entered data points automatically transferred.
Issued Caution when Switching between Data Types The Generalized HoekBrown strength model in RocData can be fit only to triaxial data, while the BartonBandis model can be applied only to direct shear data. As a result, when a user attempts to switch to one of these models while analyzing a data type that is incompatible, the program alerts the user on the inconsistency.
If the user chooses to proceed with the switch, the new data analysis dialog is opened, but without any data. The data type switch warning can be turned off if desired by selecting the “Do not show warning again” checkbox.
Data type switch warning dialog
Selection of Curve Fitting Technique
For every strength criterion in RocData there are at least two techniques available for fitting the criterion to test data. Curve fitting techniques can be selected from the dropdown menu in the upper left corner of every data analysis dialog.
31
New parameters are calculated as soon as a curve fitting method is selected from the dropdown menu.
The following table summarizes the curve fitting methods available for the different strength models:
Curve Fitting Method 

Strength Model 
Levenberg 
Simplex 
Linear 
Marquardt 
Regression 

HoekBrown (lab data / intact rock) 
X 
X 
X 
Generalized HoekBrown (rock mass / field data) 
X 
X 

BartonBandis 
X 
X 

Power Curve 
X 
X 

MohrCoulomb 
X 
X 
32
Stress Sampler
The Stress Sampler option allows the user to graphically obtain the exact stress coordinates at any point along the failure envelopes. This is done as follows:
1. Select the Stress Sampler option from the toolbar, the rightclick menu or the Analysis menu.
2. Single click the left mouse button, at any value of sigma3 (on the principal stress plot), or any value of normal stress (on the shear normal stress plot).
3. The stress coordinates will be displayed, corresponding to the value of sigma3 or normal stress, at which the mouse was clicked. A vertical dotted line will be displayed on the plots, to mark the location.
4. Alternatively, if you click and HOLD the left mouse button on either of the plots, and DRAG the mouse left or right, the stress coordinates of the failure envelopes will be continuously displayed, as you move the mouse.
5. NOTE: if the equivalent MohrCoulomb envelopes are also displayed, then the stress coordinates of BOTH the nonlinear strength, and equivalent MohrCoulomb envelopes will be displayed.
33
Instantaneous MohrCoulomb Sampler
The Instantaneous MohrCoulomb Sampler option allows the user to graphically obtain the instantaneous MohrCoulomb parameters (cohesion and friction angle), at any point along a curved failure envelope. This is done as follows:
1. Select the Instantaneous MC Sampler option from the toolbar, the rightclick menu or the Analysis menu.
2. Single click the left mouse button, at any value of sigma3 (on the principal stress plot), or any value of normal stress (on the shear normal stress plot).
3. The instantaneous MohrCoulomb envelope (tangential line) will appear on the plots, for the value of sigma3 or normal stress, at which the mouse was clicked. The instantaneous values of cohesion and friction angle will be displayed, as well as the stress coordinates. A vertical dotted line will be displayed on the plots, to mark the location.
4. Alternatively, if you click and HOLD the left mouse button on either of the plots, and DRAG the mouse left or right, the instantaneous MohrCoulomb envelope and coordinates, will be continuously displayed, as you move the mouse.
34
Exporting Data / Images
The RocData analysis results and failure envelope plots can be exported to other programs for report writing, further analysis, etc. This can be done in various ways.
• The Copy Data option in the sidebar or the Edit menu, will copy the contents of the sidebar (i.e. all input and output parameters) to the clipboard. From the clipboard, this can be pasted into a word processor or spreadsheet. This provides a handy analysis summary.
• The Copy option in the toolbar or the Edit menu, will copy an
image of the current failure envelope that is displayed, to the
clipboard. From the clipboard, this can be pasted into reports or image editing programs, etc.
• The failure envelope plots can also be saved directly to a JPEG, BMP, EMF or WMF image file, with the Export Image File option. This is available in the File menu, or the right click menu. (An image of this option is displayed below.)
One Click Export to Excel The most powerful data export feature is the following – with a single mouse click, all data AND plots can be exported to Microsoft Excel. To do this:
Select the Export to Excel toolbar button.
If you have Excel installed on your computer, the Excel program will be automatically started, and all analysis input and output data will be exported to an Excel spreadsheet.
1.
2.
3.
4.
Failure envelope plots will then be generated in Excel.
This all takes place with a single mouse click!!!
Note that the plots, which are generated in Excel, correspond to the failure envelope plot(s) that you are currently viewing. For example, if you are only viewing the shearnormal plot, then only the shear normal data and plot will be generated in Excel. If you are viewing both plots (principal stress and shearnormal), then both plots will be generated in Excel. In addition, if the equivalent MohrCoulomb envelope is displayed, this will also be plotted in Excel. Finally, note that the number of data points used to create each failure envelope is controlled in the Display Options dialog in RocData.
35
The Export option in the File submenu
36
Display Options
The appearance of the Failure Envelope plots may be customized with Display Options, zooming, and other options.
• Display Options is available in the toolbar, the View menu, and also in the rightclick menu (if you rightclick the mouse anywhere in the Failure Envelope display area).
Most of the Display Options are selfexplanatory, and it is left to the user to experiment with the options. However, we will note the purpose of the following options:
Mogi’s Line Mogi’s Line defines the ratio of major and minor effective principal stresses at which there is a transition from brittle to ductile failure. This line is simply defined by sig1/sig3 = 3.4, and is plotted as a green line on the principal stress plot, when the Mogi’s Line option is selected in the Display Options dialog.
• If the principal stress failure envelope lies ABOVE Mogi’s line, this indicates a brittle failure mode.
• If the principal stress failure envelope lies BELOW Mogi’s line, this indicates a ductile failure mode. This may occur with low values of GSI, for example.
37
Zoom Extents The Zoom Extents option will automatically scale the axes of the failure envelopes for the Generalized HoekBrown criterion, so that all possible failure envelopes, for the current value of sigci, will be visible on the plot.
This corresponds to maximum values of GSI (= 100) and mi ( = 40).
1. To demonstrate this, select the Zoom Extents option.
2. Use the interactive arrow buttons, to increase the value of GSI to 100, and the value of mi to 40.
3. Observe the failure envelope plots. When GSI = 100 and mi = 40, the extent of the principal stress envelope will correspond to the maximum extent automatically calculated by the Zoom Extents option.
You may find the Zoom Extents option useful for interactive demonstration of parameter changes on the failure envelope.
38
Project Settings
The Project Settings dialog is used to enter a project title, to select a failure criterion or to select a measurement unit.
• The Project Settings dialog is available in the toolbar, the Analysis menu.
The Project Settings dialog is selfexplanatory, and it is left to the user to experiment with the options.
39
Documentation
The calculations in the RocData program, are based on the latest version of the Generalized HoekBrown failure criterion, as detailed in the following paper (Ref. 1):
Hoek, E., CarranzaTorres, C.T., and Corkum, B. (2002), Hoek Brown failure criterion – 2002 edition. Proc. North American Rock Mechanics Society meeting in Toronto in July 2002
The program RocData incorporates all of the latest developments described in this paper.
The paper should be read by all users of the Generalized HoekBrown criterion in RocData!!! The definitions and equations for all input and output parameters in RocData, can be found in this paper, which is available as a PDF document.
Another useful document is the following (Ref. 2):
“A Brief History of the HoekBrown Failure Criterion”, by Evert Hoek.
This paper provides a brief chronological overview of the evolution of the HoekBrown failure criterion, and provides references to all of the significant papers which have been published. This is also recommended reading for all users of the Generalized HoekBrown criterion. The paper is available as a PDF document.
Finally, a set of notes by Dr. Evert Hoek, Practical Rock Engineering, (Ref. 3) is available on the Rocscience website, www.rocscience.com , in the form of a series of PDF documents. Practical Rock Engineering is recommended reading for all students of rock mechanics. Chapter 11 (Rock mass properties) should also be read by users of RocData. Note that the version of the HoekBrown failure criterion which is presented in the current version of Practical Rock Engineering, has been superceded by the latest version of the criterion (Ref. 1). However, the conceptual information is still relevant, and gives further insight into the development of the HoekBrown criterion.
40
Tutorials
Example 1 – HoekBrown Strength Envelope for a Tunnel in Undisturbed Rock (D = 0)
Consider an undisturbed in situ rock mass surrounding a tunnel at a depth of 100 meters, with the following HoekBrown classification parameters.
Enter this data in the sidebar input data area. Also, enter the following data to determine the Failure Envelope Range (sig3max), which is used to calculate equivalent MohrCoulomb parameters for the HoekBrown model.
The resulting output will be automatically displayed in the sidebar, and the failure envelopes calculated and plotted.
Notice the equivalent MohrCoulomb parameters that have been calculated:
The envelope corresponding to these parameters can be viewed on the plots, by selecting the MohrCoulomb strength envelope option, from the toolbar or the Analysis menu.
41
To get a better look at the MC envelope, let’s view only the Normal vs. Shear stress plot. Select the Normal vs. Shear Stress option from the Analysis menu or the toolbar. This will hide the principal stress plot, and show only the Normal vs. Shear stress plot, maximized in the view.
If you examine the MC envelope, you can graphically confirm the calculated values of cohesion, friction angle, and also the rock mass tensile strength sigt. The tensile strength is the negative value of normal stress, at the origin of the failure envelope.
42
Example 2 – HoekBrown Strength Envelope for a Slope Application in Highly Disturbed Rock (D = 1)
Consider a rock mass with the same basic parameters as the previous example, but in a highly disturbed slope of 100 meters height, with a disturbance factor D = 1.
Enter disturbance factor D = 1.
Enter the following data to determine the Failure Envelope Range
(sig3max):
The resulting output will be automatically displayed in the sidebar, and the failure envelopes calculated and plotted.
Notice the equivalent MohrCoulomb parameters which have been calculated:
43
Compare the equivalent MohrCoulomb parameters and envelopes, calculated in Example 1 and Example 2. This demonstrates the significant effect which the disturbance factor D, can have on the calculated rock mass strength.
For information about the Disturbance Factor, see Ref. 1 – section 7:
Estimation of Disturbance Factor D.
44
Example 3 – HoekBrown, Power Curve and MohrCoulomb Analysis of an Intact Rock Triaxial Dataset
This tutorial will involve fitting three strength models – the Hoek Brown, MohrCoulomb and Power Curve criteria  to a triaxial lab dataset for intact rock.
Start up RocData and a new analysis file will be automatically opened. If you are already in the program select the New option from the File menu or from the toolbar.
Select the Project Settings option (from the Analysis menu or the toolbar) and ensure that the measurement unit is set to MPa. Notice that the Generalized HoekBrown criterion is the default strength model.
Next select the Use Lab Data option from the Analysis menu or toolbar. This opens up the HoekBrown lab (intact) data analysis dialog. The LevenbergMarquardt technique is the default curve fitting method in the dialog.
Set the number of data points to be analyzed to 5, and enter the values shown on the figure below.
Notice that each time you hit Enter on the keyboard, sigci and mi are immediately calculated for the data that has been input.
Select the Simplex curve fitting method and observe the immediate calculation of a new fit. Next select the Linear Regression curve fitting method and view the resulting parameters.
Select the LevenbergMarquardt method again and click the OK button. This action applies the analysis results to the main view of RocData and exits the dialog.
45
The report at the top left corner of the main view displays all the relevant parameters of the analysis of the lad data. At the very end of the list, information is provided on the curve fitting method used and the value of the Residuals.
Change the sigci value displayed on the docking form to a number of your choice; note the changes in the last line of information displayed. RocData warns that the currently displayed model is not a ‘bestfit’, i.e. is not a model produced by one of the curve fitting techniques.
Go into the Project Settings dialog, select the MohrCoulomb criterion, and click OK. Immediately the data analysis dialog for the criterion is activated. The dialog contains the intact rock data previously analyzed with the HoekBrown lab data option. Click the OK button to apply the analysis results and exit from the dialog.
Next select the Power Curve strength model from the Project Settings dialog. The Power Curve data analysis dialog is opened with the input data and its corresponding fit. You may save the analysis under any file name of your choice and exit the program.
This concludes the tutorial on fitting the HoekBrown, MohrCoulomb and Power Curve strength models to an intact rock lab dataset.
46
Example 4 – Generalized HoekBrown Analysis of Field Data
This tutorial will fit the Generalized HoekBrown criterion to triaxial strength data for a rock mass.
Start up RocData and a new analysis file will be automatically opened. If you are already in the program select the New option from the File menu or from the toolbar.
Select the Project Settings option (from the Analysis menu or the toolbar) and ensure that the measurement unit is set to MPa and strength criterion set to Generalized HoekBrown.
Next select the Use Field Data option from the Analysis menu or toolbar. This opens up the HoekBrown field (rock mass) data analysis dialog.
Enter an intact rock compressive strength (sigci) of 50 MPa and a Disturbance Factor (D) of 0.5. Set the number of data points to be analyzed to 9, and enter the values shown on the figure below.
Input data for the tutorial
Select the Simplex curve fitting method and observe the immediate calculation of a new fit. Click the OK button to close the dialog and apply the analysis results to the RocData main view.
47
The default settings for the main view do not display the input data points analyzed. To view them select the Plot Test Data option from the Analysis menu or the toolbar.
Generalized HoekBrown analysis of triaxial field data
Triaxial data points are shown as points on the principal space plot of the strength envelope, and as Mohr circles on the direct shear plot.
This concludes the tutorial on fitting the Generalized HoekBrown criterion to field (rock mass) triaxial strength data.
48
Example 5 – MohrCoulomb and Power Curve Analysis of a Direct Shear Dataset
This tutorial will examine the fitting two strength models – the Mohr Coulomb and Power Curve criteria  to direct shear stress data (measured in psi units) for a soil.
Start up RocData and a new analysis file will be automatically opened. If you are already in the program select the New option from the File menu or from the toolbar.
Select the Project Settings option (from the Analysis menu or the toolbar) and set the measurement unit to psi. Change the strength model from Generalized HoekBrown to MohrCoulomb.
Project Settings for the tutorial
Next select the Use Lab Data option from the Analysis menu or toolbar. This opens up the MohrCoulomb lab data analysis dialog.
There are two radio buttons near the top of the dialog that enable users to select the type of data to be analyzed: triaxial or direct shear. Select the direct shear radio button. Set the number of data points to be analyzed to 8, and enter the values shown on the figure below.
49
Input data for the tutorial. On the shown dialog image notice the radio buttons for selecting the type of data to be analyzed.
Select the Linear Regression curve fitting method and view the resulting parameters. (For MohrCoulomb data the Linear Regression and LevenbergMarquardt methods give the same answers for practically all cases.)
Click the OK button. This action applies the analysis results to the main view of RocData and exits the dialog.
Go into the Project Settings dialog, select the Power Curve criterion, and click OK. Immediately the data analysis dialog for the criterion is activated. The dialog contains the intact rock data previously analyzed with the MohrCoulomb lab data option.
50
Power Curve fit of the sample direct shear data
This concludes the tutorial on fitting the MohrCoulomb and Power Curve strength models to a direct shear dataset.
51
References
1. Hoek, E., CarranzaTorres, C.T., and Corkum, B. (2002), Hoek Brown failure criterion – 2002 edition. Proc. North American Rock Mechanics Society meeting in Toronto in July 2002.
2. Hoek, Evert, (2002), A Brief History of the HoekBrown Failure Criterion, unpublished document.
3. Hoek, Evert, Practical Rock Engineering – An Ongoing Set of Notes, available on the Rocscience website, www.rocscience.com
52
Molto più che documenti.
Scopri tutto ciò che Scribd ha da offrire, inclusi libri e audiolibri dei maggiori editori.
Annulla in qualsiasi momento.